Struggling with warfarin? Try relying on algorithm
Sites that adhered to a simple algorithm for when to change and when not to change warfarin dose were associated with improved anticoagulation control, according to a study published online Oct. 1 in Circulation.
Harriette G.C. Van Spall, MD, MPH, of the Population Health Research Institute at McMaster University in Hamilton, Canada, and colleagues used RE-LY, a clinical trial that compared dabigatran (Pradaxa, Boehringer Ingelheim) and warfarin treatment in 6,022 nonvalvular atrial fibrillation patients who were followed for two years. The data included documentation of international normalized ratio (INR) results and decisions about the patients’ warfarin dosing.
The trial incorporated an algorithm for warfarin dosing that recommended no change for in-range results and 10 to 15 percent weekly dose changes for out-of-range INR values. Based on these data, the researchers set out to evaluate warfarin dosing practice for predicting the time in therapeutic range (TTR) and clinical outcomes. They used multilevel regression models to assess whether algorithm-consistent warfarin dosing could predict TTR and a composite outcome of stroke, systemic embolism or major hemorrhage.
They defined algorithm-consistent as within 5 percent of the dose recommended by the algorithm. The researchers could not determine if the physician used the algorithm, only that the adjustments fit the recommendations of the algorithm.
Their analyses indicated that warfarin dosing practice explained 87 percent of the between-site and 55 percent of the between-country TTR variation. Patient-level variables, on the other hand, explained only an additional 1 percent of the between-center and 20 percent of the between-country TTR variance. They found that each 10 percent increase in algorithm-consistent dosing at the site level predicted a 6.12 percent increase in TTR, and that a 10 percent increase in algorithm-consistent dosing at the site-level was associated with an 8 percent decrease in the annual rate of the composite outcome.
The results showed that clinical skill, measured as algorithm-consistency, accounted for most of the variation between countries and sites. The authors noted that in RE-LY, many countries and sites fell below the INR standard. Their results point to “sub-optimal” decisions on warfarin dose adjustments as the culprit.
“[O]ur study did not have an optimal design to establish the most effective warfarin-dosing algorithm for INR control, and further research is required to determine the best dosing strategy for slightly out-of-range INR values,” Van Spall and colleagues wrote. “This does not diminish the findings of our primary analyses, which show that adherence to a simple algorithm like the one used in RE-LY is associated with improved INR control and appears to be an easy, reliable and cost-effective way to achieve good INR control.”
They pointed out that the results likely underestimated the variation in TTR and the associations between sites and countries because the data had been collected for a clinical trial with a select patient population, a limitation also highlighted by Adam J. Rose, MD, of the Center for Health Quality, Outcomes and Economic Research at the Bedford Veteran Affairs Medical Center in Bedford, Mass.
In an accompanying editorial, Rose complimented the researchers for finding “an unusually straightforward explanation for performance variation, at least in the context of the management of warfarin.” He added that because the data included a comparison dabigatran patient group, they could tease out the relationship between TTR and outcomes.
“Some have expressed doubts about whether TTR really is in the causal pathway to outcomes, suggesting that instead, sites with higher TTR also might be delivering high-quality care in other ways,” he wrote. But if care delivery were the pathway, then the researchers would have found similar reductions in the composite outcome in the dabigatran group. “The present study therefore serves as a strong endorsement of efforts to improve TTR at the site level and thereby prevent adverse events.”
Rose added that as with any warfarin algorithm, the physician should have the ability to override the algorithm.
Harriette G.C. Van Spall, MD, MPH, of the Population Health Research Institute at McMaster University in Hamilton, Canada, and colleagues used RE-LY, a clinical trial that compared dabigatran (Pradaxa, Boehringer Ingelheim) and warfarin treatment in 6,022 nonvalvular atrial fibrillation patients who were followed for two years. The data included documentation of international normalized ratio (INR) results and decisions about the patients’ warfarin dosing.
The trial incorporated an algorithm for warfarin dosing that recommended no change for in-range results and 10 to 15 percent weekly dose changes for out-of-range INR values. Based on these data, the researchers set out to evaluate warfarin dosing practice for predicting the time in therapeutic range (TTR) and clinical outcomes. They used multilevel regression models to assess whether algorithm-consistent warfarin dosing could predict TTR and a composite outcome of stroke, systemic embolism or major hemorrhage.
They defined algorithm-consistent as within 5 percent of the dose recommended by the algorithm. The researchers could not determine if the physician used the algorithm, only that the adjustments fit the recommendations of the algorithm.
Their analyses indicated that warfarin dosing practice explained 87 percent of the between-site and 55 percent of the between-country TTR variation. Patient-level variables, on the other hand, explained only an additional 1 percent of the between-center and 20 percent of the between-country TTR variance. They found that each 10 percent increase in algorithm-consistent dosing at the site level predicted a 6.12 percent increase in TTR, and that a 10 percent increase in algorithm-consistent dosing at the site-level was associated with an 8 percent decrease in the annual rate of the composite outcome.
The results showed that clinical skill, measured as algorithm-consistency, accounted for most of the variation between countries and sites. The authors noted that in RE-LY, many countries and sites fell below the INR standard. Their results point to “sub-optimal” decisions on warfarin dose adjustments as the culprit.
“[O]ur study did not have an optimal design to establish the most effective warfarin-dosing algorithm for INR control, and further research is required to determine the best dosing strategy for slightly out-of-range INR values,” Van Spall and colleagues wrote. “This does not diminish the findings of our primary analyses, which show that adherence to a simple algorithm like the one used in RE-LY is associated with improved INR control and appears to be an easy, reliable and cost-effective way to achieve good INR control.”
They pointed out that the results likely underestimated the variation in TTR and the associations between sites and countries because the data had been collected for a clinical trial with a select patient population, a limitation also highlighted by Adam J. Rose, MD, of the Center for Health Quality, Outcomes and Economic Research at the Bedford Veteran Affairs Medical Center in Bedford, Mass.
In an accompanying editorial, Rose complimented the researchers for finding “an unusually straightforward explanation for performance variation, at least in the context of the management of warfarin.” He added that because the data included a comparison dabigatran patient group, they could tease out the relationship between TTR and outcomes.
“Some have expressed doubts about whether TTR really is in the causal pathway to outcomes, suggesting that instead, sites with higher TTR also might be delivering high-quality care in other ways,” he wrote. But if care delivery were the pathway, then the researchers would have found similar reductions in the composite outcome in the dabigatran group. “The present study therefore serves as a strong endorsement of efforts to improve TTR at the site level and thereby prevent adverse events.”
Rose added that as with any warfarin algorithm, the physician should have the ability to override the algorithm.